Luceri, L., Salkar, T. V., Balasubramanian, A., Pinto, G., Sun, C., & Ferrara, E. (2025). Coordinated inauthentic behavior on TikTok: Challenges and opportunities for detection in a video-first ecosystem. arXiv [cs.SI].
Summary
This paper develops a computational framework for detecting coordinated inauthentic behavior (CIB) on TikTok, addressing a notable gap in influence-operations research that has concentrated on text-centric platforms such as Twitter/X and Facebook. The authors argue that methods built for text-based ecosystems do not transfer cleanly to a video-first platform, and they adapt a network-based user-similarity approach to TikTok’s distinctive content and interaction structures. The core claim is that video platforms present both unique detection challenges and new opportunities, requiring methodological adaptation rather than direct importation of existing techniques.
Key Contributions
- One of the first computational frameworks tailored specifically to detecting CIB on TikTok.
- Extension of network-based CIB detection methodology from text-centric platforms to a video-first ecosystem.
- Identification of the distinctive challenges and opportunities that video platforms pose for influence-operation detection.
- Evaluation of which behavioral signals are most diagnostic of coordination in TikTok’s environment.
Methods
- Development of a computational framework for CIB detection on TikTok.
- Construction of user similarity networks based on shared behavioral or content traces, adapted from prior CIB detection literature.
- Empirical evaluation of the framework on TikTok data (specific dataset details not given in the abstract).
Findings
- The framework can surface coordinated activity on TikTok by leveraging platform-specific signals.
- Detection methods designed for text-centric platforms do not straightforwardly transfer to video-first platforms.
- Specific empirical results are not detailed in the available text.
Connections
This work extends the network-based coordination-detection tradition — using user similarity networks to surface CIB — into the video domain; that lineage is closely tied to Giglietto2020-9d8acdd7, Giglietto2022-0e951ac5, Giglietto2023-fa71a001, and the multi-platform coordination work in Minici2024-tf. Its focus on TikTok-specific signals connects it to platform studies of TikTok such as Jurg2025-ur and to broader mappings of coordinated behavior and influence operations in Gonzalez-Bailon2024-rq and Gerard2025-br.